PUF Cell Reliability Testing via Synchronized Stress Signals

The patent applies preliminary action by testing PUF cells under simulated stress conditions (thermal variations and voltage fluctuations) before final deployment. The control signal generator pre-applies stress conditions to identify weak cells, and the mask stores information about unreliable cells in advance, preventing them from affecting operational reliability.

The patent converts the harmful effect of manufacturing variations and operational stress into a benefit by using these same stress conditions to identify and filter weak PUF cells. The thermal variations and voltage fluctuations that normally degrade reliability are deliberately applied during testing to reveal unreliable cells, which are then excluded via masking, thereby improving overall system reliability.

The patent extracts and removes weak PUF cells from the identification process by generating a mask that identifies unreliable cells. The control signal generator tests each cell individually, and the mask stores information about which cells to exclude. This extraction of defective elements allows the system to use only strong cells, improving reliability without significantly reducing the usable quantity.

The patent changes the operational parameters of PUF cells by applying different stress conditions (temperature variations, voltage fluctuations, timing variations) during testing. By varying these parameters, the system identifies cells that maintain stable output across different conditions, thereby selecting cells with superior reliability characteristics while maintaining adequate quantity for identification purposes.

The patent applies universality by designing the control signal generator to perform multiple functions: it generates normal operational signals for PUF cells, generates stress test signals to identify weak cells, and controls the timing variations needed for testing. This multi-functionality reduces the need for separate dedicated testing hardware, thereby limiting the increase in device complexity while achieving reliable cell classification.

The PUF cells themselves are used to generate the identification data, and the same cells are tested using variations of their normal operational signals. The system essentially tests itself by applying controlled stress to its own components rather than requiring entirely separate testing equipment. The mask storage also uses existing memory resources within the device, allowing the system to perform self-diagnosis and self-classification with minimal additional external infrastructure.